State-Space MPC with a geometric approach for constraint handling in PEM electrolyser temperature control

This paper presents a model predictive control (MPC) with constraints for temperature control of Proton Exchange Membrane (PEM) electrolyser. The strategy is based on an MPC derived from a state-space representation that includes a disturbance model. Constraints are handled using a geometric approach known as the constraint mapping law, avoiding the need for a quadratic programming (QP) optimizer. The proposed controller offers two additional degrees of freedom in the control strategy: one for tuning the disturbance model and another for adjusting the cost function by adding a slack variable to the output constraint. Simulations and a comparative analysis highlight the advantages of the proposed controller, considering both its performance in the time domain and its computational efficiency.